SCREENING FOR DIFFERENTIALLY EXPRESSED GENES IN DENGUE INFECTION UNDER ANTIBODY DEPENDENT ENHANCEMENT CONDITIONS

SCREENING FOR DIFFERENTIALLY EXPRESSED GENES IN DENGUE INFECTION UNDER ANTIBODY DEPENDENT ENHANCEMENT CONDITIONS CHENG XUANHAO (B.Sc. (Hons), NUS) A THESIS SUBMITTED FOR THE DEGREE OF MASTER OF SCIENCE DEPARTMENT OF MICROBIOLOGY NATIONAL UNIVERSITY OF SINGAPORE 2010 Acknowledgment I would like to take this opportunity to express my sincere thanks and utmost gratitude to: Dr. Justin Wong For his guidance and support. Thank you for this opportunity. Dr Ooi Eng Eong et al For providing the humanized 3H5, original 3H5 antibody and DENV NGC strain. Professor Vincent Chow et al For providing HL-CZ cell line. Dr Sylvie Alonso et al For providing C6/36 cell line. NUS, CELS, Faculty of Medicine, Department of Microbiology, Immunology Programme, Biopolis Shared Facilities For providing the scholarship, facilities, equipments and environment for this research Staff and members of Immunology Programme, and Department of Microbiology For their direct and indirect help on the project Yati, Angeline, Xie Fei, rest of BSF staff For their constant help on microarray experiment. Clement, Janet and Zhi Hui Clement for allowing me to use Genespring in Neuro Programme. Janet and Zhi Hui for helping me with the analysis of microarray data free of charge (Courtesy of Genomax). Illumina, Genomax, Applied Biosystems, all the Bioscience Companies, and various symposiums/ conferences and seminars They had provided lots of information and technology needed for the project, and most importantly free buffets. Hazel, Daniel, Vic, Wei Bing, Chen Yu, Alvin, Gdine, and Cass- my lab mates For the operation/ proper functioning of the lab, and help in the project. All my friends For their friendship and suggestions on the project. Gordon suggested HL-CZ as a platform for DENV infection, which gave me the idea to test it out and develop it further into an ADE platform. Junji and Cher Siong for suggesting against Hsp40, or else I would have chosen Hsp40 over DDX to work on after the microarray study. i Table of contents Acknowledgment i Table of contents ii Summary vi List of Tables vii List of Figures viii List of Abbreviation ix Chapter 1 Introduction 1 1.1 Dengue virus 2 1.1.1 Classification of dengue virus 2 1.1.2 Virus structure 2 1.1.3 Virus life cycle in host cell 4 1.1.4 Pathogenesis of disease 9 1.1.5 Dengue Epidemiology 11 1.1.6 Treatment of dengue 13 1.2 Antibody Dependent Enhancement (ADE) 14 1.2.1 Mechanisms of ADE 15 1.2.2 Fc-FcR mediated entry 15 1.2.3 IgM-Complement mediated entry 18 1.2.4 Cross-binding antibody mediated entry 19 1.2.5 Role of FcγR signalling in ADE 10 1.3 Factors influencing ADE 23 ii 1.3.1 Specificity of antibody 23 1.3.2 Role of cholesterol depleting drugs on ADE 23 1.3.3 Negation of ADE by C1q 24 1.3.4 Cytokines and enzymes affect host FcγR number and function 25 1.3.5 Potential of DC-SIGN to obscure ADE 25 1.3.6 Relationship between MOI and ADE 26 1.3.7 Different cell types and virus strains affects level of enhancement 26 under ADE conditions 1.4 Types of cells used for in vitro ADE studies 28 1.4.1 Peripheral blood mononuclear leukocytes 28 1.4.2 Primary CD14+ monocytes and macrophages 29 1.4.3 Monocytic cell lines 29 1.4.4 Primary dendritic cells (DC) matured with MCM mimic 31 1.4.5 Other secondary cell lines 31 1.5 Other competing hypothesis 34 1.6 DEAD-box RNA helicases involvement in virus infections 36 1.6.1 DEAD-box helicase as a virus sensor 39 1.6.2 DEAD-box helicase in viral replication 40 1.6.2.1 DDX3 41 1.6.2.2 DDX1 42 1.6.2.3 DDX5 42 1.6.2.4 DDX24 43 1.6.2.5 DDX6 43 iii 1.6.2.6 DDX42 44 1.7 Objectives of current project 46 Chapter 2 Material and Methods 47 2.1 Cell culture 48 2.2 Virus propagation 49 2.3 Virus quantification by plaque assay 49 2.4 Infection and ADE infection of HL-CZ 50 2.6 Blocking of FcγR I and II 51 2.5 Immunostaining 52 2.7 Western Blot 53 2.8 RNA Extraction 54 2.9 Microarray 54 2.10 RT-PCR 55 2.11 qPCR 55 2.12 siRNA Transfection 57 2.13 Statistical Analysis 59 Chapter 3 Results 60 3.1 Absence of DC-SIGN on HL-CZ 61 3.2 Establishment of ADE infection in HL-CZ 62 3.3 Enhancement is maintained as long as antibody: virus ratio remains 64 3.4 Reduced fold enhancement with increasing MOI 66 3.5 Peak enhancement infection rates at peak enhancing antibody: virus 69 ratio (6 ρ g: 1pfu) for MOI 0.4 is equivalent to non-enhanced iv infection rate at MOI 0.8 (Equivalent) 3.6 Characterizing the role of FcγR on HL-CZ as a platform for ADE 72 3.7 Microarray comparison of HuADE and MoADE and its Equivalent 79 3.8 qPCR and protein expression verification of selected dead-box 81 helicases 3.9 siRNA transfection of HL-CZ knockdown both DDX mRNA and 84 protein expression at 48 hours and 72 hours respectively 3.10 Impact of DDX31 and 47 on ADE infection 87 Chapter 4 Discussion 89 Chapter 5 References 100 Chapter 6 Appendix 118 v Summary In this study, establishment of HL-CZ (a promonocytic cell line) as a new platform for study of antibody dependent enhancement (ADE) of dengue virus infection in vitro was achieved. Characteristics of HL-CZ that enable it to support ADE were also investigated. We performed microarray gene expression profiling to compare HL-CZ cells infected with dengue virus under antibody dependent enhancement (ADE) conditions versus HL-CZ cells infected to an equivalent degree but under non-enhancing conditions. We observed differential expression of several genes belonging to the DEAD-box family of RNA helicases (DDX). These observations were confirmed at a protein level by immunoblotting for these proteins in cell lysates obtained from infected cells. Subsequent experiments employing siRNA-mediated knock-down of protein expression suggested that DDX31 and DDX47 may be crucial in supporting infection of dengue virus under ADE conditions. vi List of Tables 1.1.1 1997 WHO classification of DF/DHF 10 1.1.2 2009 simplified WHO classification of DF/DHF 10 1.2 Factor affecting ADE 27 1.3 Different cells used for ADE study 33 1.4 DDX association with viruses 45 2.1 Table of primers 56 3.1 Efficacy of siRNA on DDX21, 31 and 47 mRNA expression 86 vii List of Figures 1.1 Illustration of ED homodimer 3 1.2 DENV polyprotein topology in ER membrane 6 1.3 DENV life cycle in cell host 8 1.4 Vectors that transmit DENV 12 1.5 Global distribution of DENV 12 1.6 Typical ADE infection profile to illustrate terms used in ADE studies 16 1.7A/B Illustration of Fcγ-FcγR mediated ADE 17 1.8 Illustration of IgM-complement mediated ADE 18 1.9 Illustration of cross-binding antibody mediated ADE 19 1.10 FcγR signaling in ADE 22 1.11 Conserve motif of DExD/H RNA helicase family 38 1.12 DDX as an viral RNA sensor 40 3.1 Absence of DC-SIGN on HL-CZ 61 3.2 HuADE profile at MOI 0.5 on HL-CZ 63 3.3 HuADE profile at MOI 0.3 on HL-CZ 64 3.4A Peak enhancement at various MOIs 67 3.4B Fold enhancement changes at various MOIs 67 3.5A Comparison of HuADE at MOI 0.4 with Equivalent at MOI 0.8 70 3.5B Similar infection rate of IgG control and baseline control 71 3.6A Surface expression of FcγRI, II and III on HL-CZ 73 3.6B Positive controls of Figure 3.6A using U937, k562 and NKL 74 3.7 Blocking of CD64 and 32 affects ADE on HL-CZ 76 3.8 Microarray heat map of HuADE and Equivalent treatment groups 79 3.9A qPCR verification of DDX expression 82 3.9B Western Blot verification of DDX21, 31 and 47 expression 83 3.10 Efficacy of siRNA on DDX21, 31 and 47 protein expression 86 3.11 Effect of siRNA on HuADE and Equivalent treatment groups 88 viii List of Abbreviations ADE Antibody dependent enhancement ATP Adenosine triphosphate BHK Baby hamster kidney C Capsid C1q Complement 1q C3 Complement 3 CARD Caspase activation and recruitment domain CD Cluster of differentiation CDC Center for disease control and prevention CLEC5 C-type lectin domain family 5 CRM1 Exportin 1 Ctrl Control DC Dendritic cells DC-SIGN Dendritic cell-specific intracellular adhesion molecule 3grabbing nonintegrin (also known as CD209) DDX DEAD-box RNA helicase DENV Dengue virus DF Dengue fever DHF Dengue hemorrhagic fever E Envelope ED Envelope domain FcγR Fc gamma receptor GRP/BiP Glucose-regulated protein/ Binding immunoglobulin protein HBV Hepatitis B virus HCV Hepatitis C virus ix HIV Human Immunodeficiency Virus Hsp Heat shock protein Hu3H5 Humanized 3H5 antibody HuADE Peak ADE induced by Hu3H5 IBV Infectious bronchitis virus IFN Interferon Ig Immunoglobulin IKK IκB kinase IL Interleukin IRF Interferon regulatory factors ISRE/GAS Interferon stimulated response element/ Interferon- gamma activated sequence ITAM Immunoreceptor tyrosine-based activation motif JAK Janus kinase JEV Japanese encephalitis virus kb kilobase LGP2 Library of genetics and physiology 2 M Membrane MAVS Mitochondria antiviral signaling protein MDA5 Interferon-induced helicase C domain-containing protein 1 miRISC RNA-Induced Silencing Complex loaded with miRNA miRNA Micro RNA MOI Multiplicity of infection Mo3H5 Murine 3H5 antibody MoADE Peak ADE induced by Mo3H5 MR Mannose receptor mRNA Messenger RNA x NGC New Guinea C strain NS Non-structural Nsp Non-structural protein NTP Nucleoside triphosphate PBML Peripheral blood mononuclear leukocytes pfu Plaque forming unit prM Pre-M Rab Rat sarcoma related protein Rev Regulator of virion RIG-1 Retinoic acid-inducible gene-I SARS-CoV Severe acute respiratory syndrome coronavirus siRNA Small interfering RNA SOCS Suppressor of cytokine signalling STAT Signal transducer and activator of transcription TBK TANK binding kinase TBEV Tick-borne encephalitis virus TLR Toll-like receptor TNF Tumor necrosis factor TRIF TIR-domain-containing adapter-inducing interferon-β WHO World Health Organization WNV West Nile virus YFV Yellow fever virus xi Chapter 1: Introduction 1.1 Dengue virus 1.1.1 Classification of dengue virus Dengue virus (DENV) is a flavivirus belonging to the family Flaviviridae; other members belonging to the same family are: Yellow Fever Virus (YFV), Japanese Encephalitis Virus (JEV), Hepatitis C Virus (HCV), West Nile Virus (WNV), Tick-Bourne Encephalitis Virus (TBEV) and other several encephalitiscausing viruses [Calisher et al. 1989; Blok et al. 1992]. Due to genomic sequence variation of 30-35%, DENV are categorised into four serotypes known as: DENV 1, 2, 3 and 4. Infection with one serotype does not confer protective immunity to the other three serotypes, therefore secondary or sequential infections are possible. 1.1.2 Virus structure DENV is a positive-stranded RNA virus. The virion particles are ~50nm in size with an electron dense core containing the nucleocapsid (~30nm) [Murphy et al. 1980]. DENV contain 3 structural proteins: capsid protein (C), membrane protein (M), and envelope protein (E). The virions consist of a single-stranded RNA genome encapsulated by multiple copies of the C proteins (11kDa) [Chambers et al. 1990, Ma et al. 2004, Jones et al. 2003, Chang et al. 2001]. The genomic RNA encapsulated by C protein is approximately 10.8kb long. It encodes for the 3 structural genes (C, prM and E), followed by 7 non-structural genes (NS 1, 2A, 2B, 3, 4A, 4B and 5) [Cleaves et al. 1979, Lindenbach et al. 2003]. This structure of genomic RNA and C proteins forms the nucleocapid. The nucleocapsid is in turn encapsulated by a host-derived lipid bilayer. The host-derived lipid bilayer contains 180 copies of the viral M and E glycoproteins [Kuhn et al. 2002]. 2 M proteins (8kDa) are derived from proteolytic cleavage of prM (~21kDa). prM is the precursor of M protein that consist of the M protein and a pr fragment. pr fragments are believed to function as a chaperone to stabilize E protein during viral secretion from the host endoplasmic reticulum [Konishi et al. 1993]. The main function of prM is to stabilize E protein and prevent acid-catalyzed inactivation of E protein to its fusogenic form [Guirakhoo et al. 1992, Heinz et al. 1994, Allison et al. 1995]. E protein (53kDa) consists of 3 distinct domains (EDI, II and III) [Nybakken et al. 2005]. EDI which forms a β-barrel is a central structure for EDII and III as shown in Figure 1.1. EDII contains a putative fusion peptide that is involved in the insertion into target cell membrane [Rey et al. 1995, Roehrig et al. 1998, Allison et al. 2001]. EDIII is structurally immunoglobulin-like. EDIII also contain receptor binding motifs [Crill et al. 2001]. Besides being able to block flaviviruses attachment to receptors [Modis et al. 2005], anti-EDIII antibodies can inhibit post-attachment step of virus entry. Figure 1.1: Illustration of dengue E protein homodimer structure; EDI represented in red, EDII represented in yellow/ green, and EDIII represented in blue. [Adapted from Izabela 2010] 3 1.1.3 Virus life cycle in host cell Depending on the type of host cell, DENV is known to use a myriad of different cell surface receptors to mediate infection. In mosquito cells, DENV may utilize heat-shock protein 70 (Hsp70), R80, R67 or an unidentified 45kDa surface glycoprotein for its entry into host cell. In mammalian cells, DENV uses a different set of receptors for binding and entry. Heparan sulphate [Chen et al. 1997, Germi et al. 2002, Hilgard et al. 2000], Hsp90 and 70 [Reyes-Del et al. 2005], CD14 [Chen et al. 1999], GRP78/BiP [Jindadamrongwech et al. 2004], and a 37/67-kDa high-affinity laminin receptor [Thepparit et al. 2004] have been associated with mediation of DENV binding and entry into mammalian host cell. In human myeloid cells, DENV is known to exploit certain C-type lectin receptors for infection [Fernandez-Garcia et al. 2009]. DC-specific intracellular adhesion molecule 3-grabbing nonintegrin (DC-SIGN) [Lozach et al. 2005, Navarro-Sanchez et al. 2003, Tassaneetrithe et al. 2003] , mannose receptor (MR) [Miller et al. 2008] and C-type lectin domain family 5, member A (CLEC5) [Chen et al. 2008] have been identified as receptors on human myeloid cells for DENV attachment. It is well documented that flaviviruses exploit clathrin-mediated endocytosis for cell entry, DENV is no exception [Acosta et al. 2008, Van der Schaar et al. 2008, Krishnan et al. 2007, Chu et al. 2004, Nawa et al. 1984]. After endocytosis, DENV are internalized into early endosomes which are Rab-5 positive. Membrane fusion of the DENV and the endosome take place during Rab-7 positive late endosomes stage [Van der Schaar et al. 2008]. Membrane fusion is likely to be dependent on the acidic pH of the endosome. It is also likely to vary depending on the DENV strain, as different strains have differing membrane fusion properties [Van der Schaar et al. 4 2008, Krishnan et al. 2007]. Fusion of endosomal membrane and DENV membrane results in the release of viral nucleocapsid into the cytoplasm. Once the viral RNA genome is released into the cytoplasm, the positive-sense RNA is translated into a polyprotein by ribosomes associated with the rough endoplasmic reticulum [Cylde et al. 2006]. Signal sequences within the polyprotein translocate NS1, E and part of the prM domain into ER lumen; whereas C, NS3 and NS5 remain in the cytoplasmic region. Remaining NS2A/B and NS4A/B are localized as transmembrane proteins as shown in Figure 1.2 [Perera et al. 2008]. The polyprotein is processed co- and post-translationally by viral and host proteases before viral genome replication occurs in the cytoplasm [Bressanelli et al. 2004, Modis et al. 2004]. Viral assembly is initiated near the surface of the ER where viral proteins and replicated viral RNA genome buds into the lumen of the ER forming new subviral/non-infectious immature DENV [Kuhn et al. 2002, Zhang et al. 2003]. The resultant particles are transported to the trans-Golgi network. In the trans-Golgi network, the virus particle will be post-translationally modified to reach maturity before it is released via exocytotic mechanisms [Mukhopadhyay et al. 2005]. 5 Figure 1.2: Topology of dengue viral polyprotein in ER membrane. Viral protease cleaves the polyprotein during and after translation as indicated by the arrows. [Adapted from Perera et al. 2008] 6 It is intriguing to note that certain steps of the viral assembly can be incomplete or skipped during the virus life cycle, resulting in the release of subviral particles. Capsidless subviral particles was documented in studies by Allison et al [Allison et al. 1995, Russell et al. 1980], such particles implies that encapsulation of the nucleocapsid may not be a critical step in virus life cycle [Fonseca et al. 1994, Hunt et al. 2001, Konishi et al. 2002]. Immature progeny virions were also commonly observed in vitro [Allison et al. 2003]. This is often due to incomplete cleavage of the prM by furin. Furin is an enzyme found in the trans-Golgi network and it is responsible for the cleavage of prM [Guirakhoo et al. 1992, Stadler et al. 1997]. 7 Figure 1.3: Intracellular life cycle of DENV. Diagram illustrates that DENV utilises cellular endocytosis for entry, followed by cellular translational mechanisms in the ER for viral protein synthesis. RNA replication takes place in the cytoplasm with the aid of host polymerases and NS proteins from the virus. DENV is packaged into the ER and exocytosed from the cell via the Golgi networks. [Adapted from Van der Schaar et al. 2007] 8 1.1.4 Pathogenesis of disease DENV can cause a range of mild to severe illness. The most common disease caused by DENV is known as dengue fever (DF). DF manifest as an undifferentiated febrile disease with maculopapular rash in children. Fever, headache, retro-orbital pain, myalgia, malaise, anorexia, abdominal discomfort, lymphoadenopathy and leucopenia are commonly observed symptoms among infected individuals [Watt et al. 2003]. The fever usually persists for 5 to 7 days [Fonseca et al. 2002]. Fatalities due to DF are low with proper management of symptoms. Mortality rate for a more severe form of the disease, known as dengue hemorrhagic fever (DHF), is fairly high as compared to DF [Halstead et al. 1970a]. DHF is pathophysiologically due to increased vascular permeability leading to plasma leakage. It is characterized by 1) fever, 2) hemorrhagic episodes determined by positive tourniquet test, petechiae/ecchymoses/purpura, or mucosa/gastrointestinal tract/ injection sites bleeding, 3) thrombocytopenia with 100000/mm3 or less in platelet count, 4) and evidence of plasma leakage [WHO 2010]. DHF usually last for 7 to 10 days and is more severe than DF. Mortality rate can be lowered to less than 1% if there is proper management of the circulatory fluid volume [Rothman 1999]. In severe DHF, after a few days of fever, the patient may suddenly experience a drop in body temperature followed by signs of systemic circulatory failure. The condition of the patient will spiral into a critical state of shock; death will follow within 12 to 24 hours if medical intervention is not available to recover the fluid loss [Halstead et al. 1970b]. Such cases are known as dengue shock syndrome (DSS). DSS is the most severe form of DHF, DSS is categorised as 9 GRADE III and IV DHF according to DHF classification by World Health Organization (WHO) (Refer Table 1.1). Table 1.1.1: Classification of DHF according to symptoms by WHO in 1997 WHO DHF Grading Symptoms Grade I Fever and non-specific constitutional symptoms, and positive tourniquet test and/or easy bruising. Grade II Spontaneous bleeding manifestation. Grade III (DSS) Early signs of circulatory failure, incipient shock. Grade IV (DSS) Profound shock with undetectable pulse and blood pressure. coupled with Grade I The classification of dengue is further simplified into uncomplicated and severe dengue in 2009 as WHO found that the old classification is too restrictive. DHF/DSS will be considered as severe dengue. Table 1.1.2: New and simplified classification of dengue proposed by WHO in 2009 [Adapted from WHO 2009] 10 1.1.5 Dengue Epidemiology Dengue is one of the most important mosquito-borne viral disease in the world, and is responsible for almost 50 million infections annually [Gubler et al. 2006, WHO 2010]. Up to 500000 cases of DHF and 22000 dengue associated deaths have been documented annually [WHO 2010]. In the past 50 years, incidence has increased 30 fold. 2.5 billion-of the world’s population live in areas where dengue is endemic [Solomon et al. 2001]. This means that 2 in 5 of the global population are living in areas where dengue infection is prevalent. Before 1970, only 9 countries had documented cases of DHF, since then the numbers of countries with documented cases of DHF has quadrupled [WHO 2010]. Over the years, the spread of dengue have been exacerbated by the transport of the main mosquito vector, Aedes aegypti. Global distribution of dengue highly correlates with the distribution of its main vector [Corrêa et al. 2005]. Dengue mainly affects countries in the tropical and subtropical regions, particularly in South East Asia and Latin America; several affected nations are known to be hyperendemic (co-circulation of more than 1 dengue serotype) [Jacobs et al. 2005]. Other factors that were thought to contribute to the spread of the disease includes: rapid population growth, ruralurban migration, inadequate basic urban infrastructure, and increase in amount of solid waste which provide suitable environment for Aedes larvae growth [Corrêa et al. 2005]. The mosquito vectors that are responsible for the transmission of dengue are Aedes aegypti and Aedes albopictus [CDC 2010]. 11 Figure 1.4: On the left is Aedes aegypti and Aedes albopictus is shown on the right. Both are the main vectors contributing to the spread of DENV. Aedes aegypti is a domesticated species and Aedes albopictus is a para-domesticated species, both species can be found in urbanized regions. [Adapted from CDC 2010] Figure 1.5: Global distribution of DENV. Most regions affected by dengue are located in the tropics with hot and wet climate. Tropical climate is favourable for the survival of Aedes aegypti and Aedes albopictus which contributes to the spread of the disease. [Adapted from Jacobs et al. 2005] 12 1.1.6 Treatment of dengue There is no specific antiviral drug effective in the treatment of DF/DHF. Therefore treatment is limited to the management of symptoms and supportive therapy. Mortality rate of DHF/DSS can be up to 50% high without proper medical attention. The mortality rate can be reduced to 1% if supportive care and treatment is provided promptly [Tripathi et al. 1998]. The lack of an effective antiviral treatment is compounded by the absence of an effective vaccine in the market. However prevention of dengue is possible mainly by avoiding mosquito bites and mosquito control. 13 1.2 Antibody Dependent Enhancement (ADE) The phenomenon of ADE was first described in 1930s but the first definitive study in vitro was by Hawks in 1964 [Hawks et al. 1964]. ADE is the enhancement of viral infectivity due to the presence of antibodies at either non- or sub-neutralizing conditions. Due to the presence of four DENV serotypes, anti-DENV antibodies can be homotypic (antibodies target another DENV of the same serotype as the cognate DENV) or heterotypic (antibodies target DENV of a different serotype than the cognate DENV). ADE of DENV can be caused by a few conditions. Firstly, it could be due to homotypic antibodies diluted to a concentration where it becomes subneutralising. Secondly, it could be induced by heterotypic antibodies which are diluted to a non-neutralising concentration. Lastly, it could be induced by antibodies which can cross-bind to both target DENV and the host cell surface receptors [Halstead et al. 2003]. ADE in vitro is not restricted to flaviviruses, several other viruses (eg, Ebola, Human Immunodeficiency Virus (HIV) and Severe Acute Respiratory Syndrome Coronavirus (SARS-CoV)) can also utilise ADE mechanism for infection (Takada et al. 2007, Füst et al. 1997, Kam et al. 2007). Even though definitive studies of ADE in vitro have been established for over 40 years, there has been no definitive study in vivo to prove that ADE is possible in primate mammalian host. However, there is an instance where ADE has been demonstrated in vivo in the mouse model [Zellweger et. al. 2010]. Nonetheless, ADE mechanism is widely used to explain the occurrence of DHF. A high correlation between secondary dengue infection and DHF was established by Halstead et al in 1970 [Halstead et al. 1970a], and he suggested that the anti-DENV antibodies raised during the primary infection could contribute to the severity during secondary 14 infection via ADE mechanisms. Maternal anti-DENV antibodies transferred to the infant during pregnancy were thought to be the contributing factor for DHF in new borns infected by dengue for the first time [Kilks et al. 1988]. Several studies have associated the occurrence of DHF with sequential DENV infection [Green et al. 2006, Guy et al. 2004, Halstead et al. 1970b]; and more often than not, ADE is suggested to be the cause of this association [Halstead et al. 2002]. Furthermore, high viral load in DHF is associated with increased severity of the disease and ADE is capable of inducing higher viral output per infected cell as demonstrated in vitro [Halstead et al. 2003]. 1.2.1 Mechanisms of ADE It was first proposed that enhancement of virus infectivity is contributed by an overall increase in the binding affinity of virus-antibody complex for host cells that express FcγR. Thus, the antibody-bound virus increases the probability of the virus entering the cell as compared to virus not bound to any antibody. This contributes to the higher infection rate observed in ADE. The prerequisites for ADE are: 1) The antibody must be able to bind to the virus without neutralising the virus completely, 2) the antibody used must be able to interact with host surface molecules, and 3) the host cell must possess the receptors to interact with the antibody (via Fcγ-FcγR binding for homotypic and heterotypic antibodies). 1.2.2 Fcγ-FcγR mediated entry For ADE of DENV, a heterotypic antibody which is cross-reactive to the target virus can be used to induce enhancement. However it must be noted that even though the antibody is heterotypic, at a high enough concentration it could still 15 neutralise the virus. Therefore, the heterotypic antibody must be diluted to a subneutralising concentration before it could induce enhancement [Takada 2003]. Alternatively, a homotypic antibody can be use in place of the heterotypic antibody. Likewise, the concentration of the homotypic antibody in use is of a concern. It must be a level which it is non-neutralising but still at a level high enough to induce the enhancement effect [Morens 1987] (Refer Figure 1.6). Figure 1.6: Diagram showing the relationship between infection rate and antibody concentration in ADE. Neutralization occurs at higher antibody titres, neutralisation is lost with subsequent antibody dilutions and enhancement peak at an optimal antibody dilution. When antibody is diluted beyond peak enhancement, infection rates starts to decrease till it coincides with that of control. [Adapted from Halstead et al. 2003] Both heterotypic and homotypic antibody mediated ADE have similar mechanisms of enhancing infection. Usually the antibody used is of IgG subclass. Relying on the high affinity binding of the Fcγ portion of the antibody to the FcγR on the host cell, interaction of the virus and the host receptor which mediates viral entry is enhanced. This enhancement increases overall infection rate. Fcγ portion of the antibody-virus complex could also facilitate entry of virion via FcγR mediated endocytosis (refer Figure 1.7). FcγRI is known to have high affinity to IgG and is 16 one of the receptors involved in ADE [Kontny et al. 1988]. FcγRII which has a lower affinity for IgG has been known to be involved in ADE [Littaua et al. 1990, Rodrigo et al. 2006]. Fcγ-FcγR mediated mechanism has been widely studied, because of possible implications to DHF in sequential dengue infection. Figure 1.7A and B: Illustration of Fcγ-FcR mediated ADE. A) At high antibody titre, antiDENV antibody binds to surface of the virion. Steric hindrance from the antibody prevents binding of virus to host surface receptor. B) At enhancing antibody titre, antibody provides the steric hindrance that impedes viral entry, viral ligands can still interacts with the receptor on host cell for binding and entry. In addition the Fcγ portion of the antibody acts as a co-receptor to enhance viral binding and entry. [Adapted from Tadaka et al. 2003] 17 1.2.3 IgM-Complement mediated entry Even though most in vitro studies utilized IgG for ADE, there is one instance where IgM was able to induce ADE in WNV as well [Cardosa et al. 1983]. Instead of the Fc γ -Fc γ R binding mechanism, it was postulated that classical pathway activation of complement by IgM results in attachment of complement protein C3 fragment to WNV. The attachment of C3 fragment on the virus mediated ADE via its binding to complement receptor 3 on the host cell surface (refer Figure 1.8). Figure 1.8: Illustration of IgM-complement mediated ADE. IgM bound to DENV results in activation of classical complement pathway. This in turn results in the attachment of C3 fragment to the virion. The attached C3 interacts with complement receptor on host surface. It was postulated that the C3-complement receptor interactions function as a co-receptor to enhance DENV binding and entry. [Adapted from Cardosa et al. 1983] 18 1.2.4 Cross-binding antibody mediated entry Besides heterotypic and homotypic antibodies, antibody that cross bind the virus and host surface molecules could also induce ADE. This mechanism is not as well studied as Fcγ-FcγR mechanism. However, there are 2 independent studies demonstrating such mechanism is possible in vitro. Conjugation of anti-DENV E/prM antibody and anti-β 2 microglobulin antibody yields a chimeric antibody that is capable of cross binding DENV to β2 microglobulin of host cell. Such chimeric antibody was shown to be capable of inducing ADE even in FcγR-/- host cells [Mady 1992] (refer Figure 1.9). Using anti-prM IgG and in the absence of complement, Huang et al manage to induce ADE in FcγR-/- host cells, and it was discovered that the anti-prM IgG was able to cross-bind to Hsp60 on the host surface [Huang et al. 2006] . While chimeric and cross-binding antibodies demonstrated that nonspecificity of the antibody can also enhances viral infection in vitro, there is no definitive evidence indicating that such a mechanism is a possible contributing factor to dengue severity in vivo as there has been an absence of proof that antibodies raised during DENV infection can cross-bind to host cell membrane proteins. Figure 1.9: Proposed mechanism of bi-specific antibody induced ADE. Chimeric bi-specific antibody which consists of 2 different Fab fragments conjugated chemically can induce ADE in FcγR-/- cells. [Adapted from Mady et al. 1992] 19 1.2.5 Role of FcγR signalling in ADE Beside the traditional view of enhanced virus uptake via the Fcγ-FcγR mechanism in ADE, it was also proposed that ligation of the Fcγ portion of the antibody-virus complex with the host FcγR might result in FcγR signalling within the host cell. Both FcγR signalling together with the enhanced uptake of the virus could be the contributing factors to an overall increase in viral output per cell as observed in most ADE infection in vitro. This novel postulation explains ADE driven immunopathology such as the increased in viral load in DHF patients. Both FcγR I and IIA were identified as the FcRs responsible for mediation of ADE. Even though immunoreceptor tyrosine-based activation motif (ITAM) is found only on the cytosolic domain of FcγRIIA, it is still well documented that both FcγR I and IIA utilises ITAM for cell signalling [Abdel Shakor et al. 2004, Huang et al. 1992, Indik et al. 1991, Kwiatkowska et al. 2003, Sobota et al. 2005]. ITAM is known to have an important role in FcγRIIA mediated ADE. It was demonstrated that ADE was completely abrogated when there is an absence or mutation of the Fcγ RIIA ITAM domain [Moi et al. 2009]. Fc γ RIIIA also utilizes ITAM for its intracellular signalling but there is no definitive studies showing that FcγRIIIA is involved with ADE. Ligation of Fc to FcγR during ADE has consequences that may influence intracellular anti-viral response of the host cell. In vitro studies show that ligation of the FcγR during ADE induces IL-10 production [Mahalingam et al. 2002]. IL-10 is recognised as a key cytokine in anti inflammatory and immunosuppressive responses. 20 IL-10 upregulates suppressor of cytokine signalling 3 (SOCS3) which is responsible for repression of IFNα induced gene activation in monocytes [Ito et al. 1999, Song et al. 1998]. During DHF, TNFα level in the circulatory system is elevated and it is thought to be one of the contributing factors for plasma leakage observed in DSS [Cardier et al. 2005]. IL-10 is known to suppress TNFα via SOCS3 upregulation in vitro, this seem to contradict the hypothesis of ADE being the underlying mechanism for DHF. Therefore, it was speculated that local autocrine of IL-10 early in the infection contributes to the peak viraemia. The elevation of TNFα occurs during later stages of DHF (after viraemia had peaked), and by then systemic IL-10 level had already dropped [Green et al. 1999, Suhrbier et al. 2003]. The Fc γ R signalling during ADE not only induces IL-10 production in macrophages, it also suppresses IL-12, IFN-γ and IFN-α/β [Chareonsirisuthigul et al. 2007, Yang et al. 2001]. These cytokines are known for mediating both innate and adaptive intracellular anti-viral responses. Suppression of IL-12, IFN-γ and IFN-α/ β result in downregulation of STAT-1 and IRF-1. STAT-1 and IRF-1 are transcription factors for iNOS gene which is responsible for nitric oxide production. Overall reduction in nitric oxide levels during ADE renders the host cell more permissive to viral replication. Therefore, it could contribute to the higher viral output per ADE infected cell [Chareonsirisuthigul et al. 2007, Yang et al. 2001]. 21 Figure 1.10: Intracellular signalling triggered by Fc-FcR mediated ADE. ADE induces FcR signalling which can result in upregulation of IL-10 and reduced IL-12, IFN-γ and IFN-α/β. Solid lines indicate pathways enhanced by ADE. Dotted lines indicate pathways inhibited by ADE. [Adapted from Chareonsirisuthigul et al. 2007] 22 1.3 Factors influencing ADE As discussed in earlier paragraphs, subclass of the antibody used in ADE infection can have an impact on the mechanism by which enhancement occurs, concentration of the antibody used is also of concern. Besides antibody subclass and concentration, there are other factors which can influence the level of enhancement in ADE infections. 1.3.1 Specificity of antibody Most documented in vitro studies of ADE of DENV utilises antibody that target E protein of the viruses. However, there have been recent reports that anti-prM antibody is able to induce enhancement in vitro [Huang et al. 2006, Dejnirattisai et al. 2010]. Anti-prM antibodies were known to be highly cross reactive among the 4 serotypes, and were unable to fully neutralise DENV even at high concentration. Unlike most anti-E antibodies which show neutralisation at higher concentration, antiprM fails to neutralise DENV at high concentration of 30µg/ml. Not only did the antiprM fail to neutralise, it enhanced infection by more than 3 fold (from 20 to 70%) at 30µg/ml concentration [Dejnirattisai et al. 2010]. Dejnirattisai et al (2010) demonstrated that the specificity of the antibody used in ADE is an important parameter that influences enhancement, less specific and highly cross reactive antibodies such as anti-prM antibodies are more prone to enhancement induction. 1.3.2 Role of cholesterol depleting drugs on ADE A recent study has shown that the level of infection enhancement by ADE infection of differentiated U937 monocytic cell lines with DENV was dependent on 23 the presence of cholesterol and cholesterol-rich membrane micro-domains on the host cell. Association of FcγR with lipids rafts upon IgG binding was known to be crucial for Fc γ R receptor signalling [García-García et al. 2007, Kono et al. 2002, Kwiatkowska et al. 2001]. Drugs which deplete cholesterol and cholesterol-rich membrane micro-domains can disrupt lipid raft integrity [Reyes-del Valle et al. 2005], thereby having an adverse effect on ADE infection of the host cell. Nystatin, filipin and β-methyl cyclodextrin significantly lower ADE infection rate of differentiated U937 in vitro by disrupting the integrity of lipid rafts [Henry et al. 2010]. This drug induced reduction in ADE infection rate can be reversed by the supplementation of bovine fetal serum. Bovine fetal serum supplement replenishes the cholesterol that is needed for the formation of lipid rafts and proper ADE mechanism to occur [Henry et al. 2010]. 1.3.3 Negation of ADE by C1q Complement proteins such as C1q could negate the enhancing effect observed in ADE infection as well [Modis et al. 2004]. Presence of complement in Fcγ-FcγR mediated ADE lowers the enhancement of infection significantly. Presence of C1q lowers the peak enhancement of ADE mediated by IgG greatly. This reduction effect is more profound with IgG subclasses, such as IgG2a. IgG2a is known to bind to C1q avidly. Given that C1q is a large multimeric protein and its binding site is in close proximity to that of FcγR binding site. It was suggested that C1q restriction of ADE is contributed by the blocking of Fcγ-FcγR interaction when C1q binds to the IgG involved [Mehlhop et al. 2007, Yamanaka et al. 2007]. Exact mechanism of the C1q effect on ADE is yet to be elucidated. 24 1.3.4 Cytokines and enzymes affect host FcγR number and function Modulation of both function and expression of FcγRs on host cells are shown to have a great impact on ADE infection of DENV in vitro. Cytokines and enzymes that could up-regulate the number of FcγR on the host cell could potentially augment ADE mechanism, thereby enhancing infection rate. U937 cells treated with IFNγ is known to have an increased in peak enhanced infection rate (from 25% to 60%) under ADE condition. This was later proven to be contributed by the stimulation of U937 by the cytokine, causing an increase in number of FcγRI expressed per cell [Kontny et al. 1988]. Enzymes such as neuraminidase were also capable of modulating the expression and function of FcγRs. K562 erythroleukemic cell line pre-treated with neuraminidase was shown to be more permissive to the enhancing effect of ADE mediated infections. The enzyme was demonstrated to increase the expression of Fcγ RII in K562 and also increasing the affinity of FcγRII [Mady et al. 1993]. 1.3.5 Potential of DC-SIGN to obscure ADE Other than the expression intensity of FcγRs on the host cell, presence of other receptors may also affect the infection enhancing phenomenon of ADE mechanism. Despite the fact that DC-SIGN is a receptor which facilitates DENV entry into DC [Navarro-Sanchez et al. 2003, Tassaneetrithe et al. 2003], expression of DC-SIGN negatively impacts the effect of ADE [Boonak et al. 2009]. Transduced K562 and U937 cells that express high levels of DC-SIGN were not able to support ADE infection of DENV at all. This could be attributed to the preferential uptake of 25 the virus by DC-SIGN, thus rendering FcγR-mediated entry non-operational. High level of DC-SIGN obscuring ADE was also reported in other flavivirus infection models [Goncalvez et al. 2007, Pierson et al. 2007]. 1.3.6 Relationship between MOI and ADE The amount of virus used for the infection can have an impact on ADE as it influences the baseline infection rate. With a higher baseline infection rate there will be less room for enhancement. Therefore, it was generally observed that high multiplicity of infections (MOIs) obscure enhancement. MOI refer to the amount of viruses (in pfu/ml) exposed to each host cell during the viral absorption step of invitro infection. ADE infection in peripheral blood mononuclear cell (PBMC) was observed at MOIs of 0.001 to 0.1. However, this enhancing effect was lost when MOI was increased beyond 1[Halstead 2003]. This is probably contributed by the high baseline infection rate due to the high MOI. 1.3.7 Different cell types and virus strains affect level of enhancement under ADE conditions Different virus strains and cell types also affect the ability of DENV to undergo ADE infection. Myeloid cell lines that support ADE in vitro have different capacity to support ADE. Besides demonstrating that infection of human cell by DENV is modulated by different cell types and virus strains, Diamond et al. also demonstrated their impact on ADE of DENV infection. Using monoclonal antibody 4G2 to induce ADE infection, U937 was demonstrated to be more permissive than THP-1 monocytic cell line across 4 different strains of DENV2. Comparing infection rates between different strains on U937 alone; DENV 2 N9622 strain was unable to induce any significant enhancement whereas DENV 2 16681 had a 23% increment in 26 infection rate due to the enhancing antibody. The difference in virulence of the strain and the susceptibility of different cell types to different DENV strains clearly affected the rate of enhancement [Diamond et al. 2000]. Table 1.2: Summary of parameters and factors that could influence infection rate in ADE of DENV infection Factors affecting in vitro ADE Effects Specificity of antibody used Less specific antibodies are more likely to induce ADE Nystatin, filipin and β-methyl cyclodextrin (cholesterol depleting drugs) Depletion of cholesterol disrupts ADE C1q Presence of C1q negates ADE IFNγ and neuraminidase (cytokines and enzymes that augments the effects of Fc γR) Augments and further increases enhancement induced by ADE DC-SIGN Presence of DC-SIGN obscures ADE MOI Depends, generally ADE is loss at higher MOIs Cell types and virus Depends 27 1.4 Types of cells used for in vitro ADE studies The pre-requisite for hosting ADE infection in cell is that the host must possess either FcγRI, II or both the FcγRs. There are several cell types that had been proven in previous studies to be able to support ADE of DENV infection. 1.4.1 Peripheral blood mononuclear leukocytes ADE of DENV was first demonstrated by Halstead et al. in 1977, it was established with peripheral blood mononuclear leukocytes (PBMLs) from primate origin. In their study, they did not manage to identify the exact leukocyte subpopulation in the peripheral blood that was responsible for ADE of dengue infection [Halstead et al. 1977]. A separate study by Yang et al. also managed to establish ADE of DENV infection in PBML [Yang et al. 2001]. PBML was commonly used in the past as a platform to study the effects of ADE because of its susceptibility to DENV infection. There is a subpopulation of the cells in PBML that possess at least FcγRI or II, given that it is a primary cell type, and it better represents in vivo conditions than secondary cell types [Ross et al. 2010]. However in both studies, the exact leukocyte subpopulation in the peripheral blood that is responsible for ADE of DENV infection was not identified. Contribution of confounding by-stander cells that do not support ADE may interfere with the observation made during the study of effects induced by ADE. Therefore, there is a need for a more homogeneous cell type that can be used as a suitable platform for ADE studies. 28 1.4.2 Primary CD14+ monocyte and macrophages It is widely recognised that cells from the myeloid linage are potential hosts for dengue infection in vivo. Furthermore, myeloid cell types possess at least one of the Fc γ R required; therefore it is likely that myeloid cells such as primary monocytes could be used to study ADE in vitro. Indeed, CD14+ monocytes isolated from PBMLs were susceptible to ADE of dengue infection [Kou et al. 2008]. Primary monocytes possess all 3 subclasses of FcγRs [Halstead et al. 2003]. Both FcγRI and II contribute to ADE of DENV infection in primary monocytes [Kou et al. 2008]. Likewise for primary macrophages, macrophages extracted from the spleen were able to host dengue infection in the presence of enhancing titres of DENV-immune serum [Blackley et al. 2007]. Both primary monocytes and macrophages originate from myeloid cell linage and both support ADE in vitro. These make them an ideal platform for study of ADE in vitro as they represent in vivo conditions more closely than secondary cell lines. However, there is a major drawback when it comes to using primary cell types. Although they support ADE of DENV infection, they are not as permissive to non-enhanced DENV infection as compared to their secondary cell line counterparts [Halstead et al. 1981]. After 48 hours, less than 5% of the primary monocytes stained positive for E protein after exposure to DENV 2 16681 at a MOI of 5. Under the same conditions, only less than 1% of the primary splenic macrophages are detected as positive for DENV infection [Blackley et al. 2007]. 1.4.3 Monocytic cell lines THP-1 is a monoyctic secondary cell line that resembles primary monocytes and can support ADE. Just like monocytes, it utilises both FcγRI and II for ADE of 29 DENV infection [Chareonsirisuthigul et al. 2007, Paradkar et al. 2010, and Diamond et al. 2000]. Unfortunately it has the same drawback as primary cell types. ADE independent infection of THP-1 with DENV2 16681 at MOI of 10 only yields a mere 0.2% infection rate after 96 hours of infection [Diamond et al. 2000]. Other monocyte-like secondary cell lines, such as, U937 and K562 also support ADE mechanism [Diamond et al. 2000, Kontny et al. 1988, Littaua et al. 1990, Huang et al. 2006, Henry et al. 2010, Guy et al. 2004, Konishi et al. 2010]. U937 like its primary counterpart is also difficult to infect without the use of enhancing antibodies. Using DENV 2 NGC at a MOI of 5, only 2% of the U937 are stained positive for dengue antigen. Other strains (such as DENV2 16681, C0477 and K0049) yield an even lower rate of infection ( 104 pfu/ml CD14+ monocytes CD14+ splenic macrophages THP-1 Primary human monocytes Multiple, no definitive study FcγRI and II DENV 2 16681 5 3.54% 10.04% Kou et al., 2008 Primary human macrophage FcγRI and II DENV 2 16681 5 0.15% 11% Blackley et al., 2007 Secondary human monocyte leukemic cells Secondary human monocytic cell line Secondary human erythroleukemia cell line PBML-derived primary human cell Secondary human B cell line Secondary human mast cell-like cell line Secondary human promonocytic cell line FcγRI and II DENV 2 16681 10 [...]... MAVS Mitochondria antiviral signaling protein MDA5 Interferon-induced helicase C domain-containing protein 1 miRISC RNA-Induced Silencing Complex loaded with miRNA miRNA Micro RNA MOI Multiplicity of infection Mo3H5 Murine 3H5 antibody MoADE Peak ADE induced by Mo3H5 MR Mannose receptor mRNA Messenger RNA x NGC New Guinea C strain NS Non-structural Nsp Non-structural protein NTP Nucleoside triphosphate... the antibody can also enhances viral infection in vitro, there is no definitive evidence indicating that such a mechanism is a possible contributing factor to dengue severity in vivo as there has been an absence of proof that antibodies raised during DENV infection can cross-bind to host cell membrane proteins Figure 1.9: Proposed mechanism of bi-specific antibody induced ADE Chimeric bi-specific antibody. .. 1999], GRP78/BiP [Jindadamrongwech et al 2004], and a 37/67-kDa high-affinity laminin receptor [Thepparit et al 2004] have been associated with mediation of DENV binding and entry into mammalian host cell In human myeloid cells, DENV is known to exploit certain C-type lectin receptors for infection [Fernandez-Garcia et al 2009] DC-specific intracellular adhesion molecule 3-grabbing nonintegrin (DC-SIGN)... cholesterol-rich membrane micro-domains can disrupt lipid raft integrity [Reyes-del Valle et al 2005], thereby having an adverse effect on ADE infection of the host cell Nystatin, filipin and β-methyl cyclodextrin significantly lower ADE infection rate of differentiated U937 in vitro by disrupting the integrity of lipid rafts [Henry et al 2010] This drug induced reduction in ADE infection rate can be reversed... is capable of inducing higher viral output per infected cell as demonstrated in vitro [Halstead et al 2003] 1.2.1 Mechanisms of ADE It was first proposed that enhancement of virus infectivity is contributed by an overall increase in the binding affinity of virus -antibody complex for host cells that express FcγR Thus, the antibody- bound virus increases the probability of the virus entering the cell as... influencing ADE As discussed in earlier paragraphs, subclass of the antibody used in ADE infection can have an impact on the mechanism by which enhancement occurs, concentration of the antibody used is also of concern Besides antibody subclass and concentration, there are other factors which can influence the level of enhancement in ADE infections 1.3.1 Specificity of antibody Most documented in vitro... surface of the virion Steric hindrance from the antibody prevents binding of virus to host surface receptor B) At enhancing antibody titre, antibody provides the steric hindrance that impedes viral entry, viral ligands can still interacts with the receptor on host cell for binding and entry In addition the Fcγ portion of the antibody acts as a co-receptor to enhance viral binding and entry [Adapted from... prevent acid-catalyzed inactivation of E protein to its fusogenic form [Guirakhoo et al 1992, Heinz et al 1994, Allison et al 1995] E protein (53kDa) consists of 3 distinct domains (EDI, II and III) [Nybakken et al 2005] EDI which forms a β-barrel is a central structure for EDII and III as shown in Figure 1.1 EDII contains a putative fusion peptide that is involved in the insertion into target cell membrane... not bound to any antibody This contributes to the higher infection rate observed in ADE The prerequisites for ADE are: 1) The antibody must be able to bind to the virus without neutralising the virus completely, 2) the antibody used must be able to interact with host surface molecules, and 3) the host cell must possess the receptors to interact with the antibody (via Fcγ-FcγR binding for homotypic and... vaccine in the market However prevention of dengue is possible mainly by avoiding mosquito bites and mosquito control 13 1.2 Antibody Dependent Enhancement (ADE) The phenomenon of ADE was first described in 1930s but the first definitive study in vitro was by Hawks in 1964 [Hawks et al 1964] ADE is the enhancement of viral infectivity due to the presence of antibodies at either non- or sub-neutralizing ... Establishment of ADE infection in HL-CZ 62 3.3 Enhancement is maintained as long as antibody: virus ratio remains 64 3.4 Reduced fold enhancement with increasing MOI 66 3.5 Peak enhancement infection rates... signaling protein MDA5 Interferon-induced helicase C domain-containing protein miRISC RNA-Induced Silencing Complex loaded with miRNA miRNA Micro RNA MOI Multiplicity of infection Mo3H5 Murine... profiling to compare HL-CZ cells infected with dengue virus under antibody dependent enhancement (ADE) conditions versus HL-CZ cells infected to an equivalent degree but under non-enhancing conditions